FDA’s Recent Updates to General Wellness and Clinical Decision Support Policies – What They Mean (and What They Don’t)

On January 6, 2026, the U.S. Food and Drug Administration (“FDA”) released updated versions of two key guidance documents – General Wellness: Policy for Low Risk Devices and Clinical Decision Support Software.  The issuance of these two updated policies – and on the same day – generated significant commentary across industry stakeholders, with some positing that they hearkened notable de-regulation in the software/medical device spaces.  As detailed below, while the updated guidance documents do provide a measure of greater regulatory flexibility in certain areas, they are not, in fact, wholesale revisions of FDA’s pre-existing policies.

General Wellness Policy

This guidance document explains FDA’s policy of enforcement discretion – not enforcing medical device regulatory requirements – for products that are intended for certain health-related uses and pose low risk.  This past Wednesday (February 11), FDA hosted a virtual town hall to explain what has changed, and what has remained the same, in the updated guidance, which supersedes the version issued on September 27, 2019. The town hall itself lasted less than 30 minutes, bolstering the fact that much of the policy remains unchanged.

FDA emphasized that the policy’s key principles have not been modified.  It still encompasses two categories of intended use:

1. Those related only to general wellness without a link to any particular disease or condition; and
2. Those related to sustaining or enhancing functions associated with general health with reference to a chronic disease or condition, but only where it is well-established that healthy lifestyle choices may help reduce the risk of, or live well with, the disease or condition.

The first category encompasses claims pertaining to physical fitness, weight management, self-esteem, sleep hygiene, and similar lifestyle or general health effects.  The second category encompasses claims pertaining to, for example, helping a user live well with anxiety, reduce the risk of high blood pressure, or live well with type 2 diabetes, but still shop short of indicating a measurable clinical outcome (e.g., treatment or prevention).

To fall under general wellness enforcement discretion, the product must also be deemed not to pose notable risk to patients/users, which FDA defines as being non-invasive (not penetrating the skin or mucous membranes), not implanted, and not involving a technology that may pose a risk without the application of specific regulatory controls (e.g., lasers, radiation).  Importantly, because FDA will not even assess whether a product categorized as “general wellness” is or is not a medical device under the Federal Food, Drug and Cosmetic Act, the updated guidance still emphasizes that inclusion under the policy does not establish that a product has been shown to be safe or effective for its intended use.

The primary change from 2019 is with respect to products that are intended for non-invasive (e.g., optical) sensing to estimate, infer, or output physiologic parameters. FDA clarified at the town hall that the changes here are intended to support low-risk innovation.  Specifically, under the updated guidance, products that use non-invasive sensing for certain outputs (e.g., blood pressure, oxygen saturation, blood glucose, heart rate variability) may fall under the policy if the output is intended only for general wellness and not for a medical or clinical purpose, the labeling/functionality are appropriately limited so as not to prompt or guide specific clinical action or medical management, and output parameters that resemble ones used clinically are validated. 

Wearables, which have become increasingly popular and the subject of much public debate, are most impacted by this policy revision.  FDA’s recent battle with the company WHOOP centered on whether its WHOOP strap, a wrist-worn device that monitors and provides insights on various physiological signals, could output “Blood Pressure Insights” without a premarket authorization.  While the company maintained that the daily estimations of systolic and diastolic blood pressure fall under the general wellness policy, FDA insisted – culminating in a July 14, 2025 Warning Letter – that such estimations are “inherently associated” with a diagnosis of hypo- or hyper-tension, and thus are a medical device intended use.  The agency cited precedent as support for its position.  Non-invasive Blood Pressure Measurement Systems (Product Code DXN) are defined by regulation as devices that provide “a signal from which systolic, diastolic, mean, or any combination of the three pressures can be derived through the use of transducers placed on the surface of the body”.  Under that code, FDA has authorized other blood pressure measurement products intended to measure or estimate a user's blood pressure without explicit reference to a diagnosis of hypo- or hypertension, based on the measurement's inherent association with those conditions.  In July 2025, for example, FDA cleared the Aktiia G0 (now “Hilo Band”) as the first over-the-counter cuffless blood pressure monitor.  So, while the recent policy update may be welcomed by many in the wearables space who have not yet gone through FDA review, it may well lead to frustration among those who see themselves as having gone through the more rigorous process for similar or comparable products.

The updated policy also formally extends enforcement discretion to recognition or notification of an “out of bounds” result, as long as the product does not clinically qualify such output.  In other words, FDA notes that products that recommend a user sees a health care practitioner when the output falls outside of the established “general wellness” range may be considered general wellness products, as long as the output is not characterized (e.g., as “abnormal”), does not identify or name a specific disease/condition, and does not otherwise directly impact clinical management (e.g., by including clinical thresholds or treatment recommendations, or providing ongoing alerts or monitoring meant to manage a disease/condition).

Most of the examples provided in the guidance are unchanged from 2019, bolstering the conclusion that the policy is in many ways not fundamentally different than before. Such examples include software that monitors and records daily energy expenditure and workouts to provide awareness of one’s exercise to improve/maintain good cardiovascular health, or a product intended to mechanically exfoliate – without penetrating – the skin to make it smoother.  In one example, FDA added monitoring of oxygen saturation to the monitoring of pulse rate during exercise and hiking as falling under general wellness. 

Two new examples in this iteration help illuminate where there is now additional flexibility:

• A wrist-worn wearable product intended to assess activity and recovery that outputs multiple biomarkers, including hours slept, pulse rate, and blood pressure. FDA explains that this intended use relates to general wellness and does not refer to a specific disease or condition, and the technology for monitoring these biomarkers does not pose a risk to safety if specific regulatory controls are not applied.  Thus, as long as the product provides validated values for blood pressure and its claims do not imply use in a medical or clinical context, it falls under general wellness.
• A wearable product intended for elite athletes to monitor parameters such as electrolyte balance, lactate, and hemoglobin by displaying values from cleared devices or ranges from the product’s own optical sensor. Because the claimed use is limited to exercise/fitness contexts and the technology for measuring these physiologic parameters does not pose a safety concern, this is a general wellness product as long as the outputs are validated values.

Clinical Decision Support Software Policy

The updated Clinical Decision Support Software guidance (superseding the version released on September 28, 2022), while also not entirely revamped, does provide some notable new leniency in FDA’s assessment of whether health-related software warrants regulatory oversight.  The added flexibility is likely to prove valuable to numerous software manufacturers, including those making AI-based chatbots – which have become more common of late and were the subject of FDA’s second Digital Health Advisory Committee meeting in November 2025.

The general criteria for a product to qualify as non-device clinical decision support software (“non-device CDS”), which are derived from statutory updates made under the 21^st Century Cures Act, remain the same.  Non-device CDS includes software functions that satisfy all four of the below factors:

1. Not intended to acquire, process, or analyze a medical image or a signal from an in vitro diagnostic device, or a pattern or signal from a signal acquisition system (“Criterion 1”);
2. Intended to display, analyze, or print medical information about a patient or other medical information (e.g., peer-reviewed clinical studies, clinical practice guidelines) (“Criterion 2”);
3. Intended to support or provide recommendations to a health care professional about prevention, diagnosis, or treatment of a disease or condition (“Criterion 3”); and
4. Intended to enable the health care professional to independently review the basis for its recommendations, such that they need not rely primarily on such recommendations to make a clinical diagnosis or treatment decision regarding an individual patient (“Criterion 4”).

In the past, Criterion 4 has been key in distinguishing between what software products used in clinical decision-making are and are not subject to FDA regulation.  The agency has stressed that a level of transparency and explainability are needed for non-device CDS, so that the clinician using the software can understand – and “vet” – how the software generates its outputs.  To that end, FDA has laid out quite granular expectations of the types of information that manufacturers should provide through/about their software for it to meet non-device CDS Criterion 4.  Otherwise, FDA is concerned about automation bias, where a clinician is likely to rely on software outputs without confirming or critically evaluating them based on their clinical judgment/expertise. The emphasis on transparency is maintained in the 2026 guidance, with FDA further accentuating the need to present information to satisfy Criterion 4 with usability in mind – prioritizing decision-relevant details and avoiding overload to enhance the clinician’s ability to interpret the output. 

The 2022 iteration of the guidance markedly indicated that outputting a specific treatment directive (as opposed to a selection of option), a time-critical recommendation, or a risk probability score all would lead software to fail non-device CDS Criterion 3.  This made significant waves, with many manufacturers of software that generated a risk score or priority recommendation of some kind recognizing that products they had offered as unregulated may in fact be considered medical devices. 

The 2026 guidance notably revises this interpretation of Criterion 3. First, the limitation on singular treatment directives is softened.  FDA now explicitly recognizes that there is sometimes only one clinically appropriate recommendation.  Accordingly, despite such software not formally satisfying Criterion 3, so as not to require the provision of unreasonable suggestions to avoid regulation, FDA plans to exercise enforcement discretion for CDS that produces only a single clinically appropriate recommendation, if all other non-device CDS criteria are met.  Based on this change, two examples of software products categorized as device CDS solely based on having a specific diagnostic/treatment output which were included in the 2022 guidance have been removed, as they would now presumably be considered to fall under enforcement discretion:

1. software that identifies a patient with possible diagnosis of opioid addiction based on analysis of patient-specific medical information, family history, prescription patterns, and geographical data; and
2. software that recommends a specific, FDA-approved chemotherapeutic agent for a patient based on analysis of patient diagnosis and pathologist-confirmed biopsy results.

Along the same lines, the updated guidance recognizes that a risk probability score may represent a case where there is only one clinically reasonable output and thus may fall under enforcement discretion, if all other non-device CDS criteria are met.  Furthermore, while time-critical decision-making still typically precludes qualification as non-device CDS, the 2026 guidance now discusses this factor – along with the level of automation – under Criterion 4 instead of Criterion 3, commenting that these parameters impact whether software allows a clinician to independently review its output.  

Finally, the updated guidance also provides useful clarification with respect to non-device CDS Criterion 1.  FDA elaborates that a continuous, near-continuous, or streaming measurement would fail this criterion, whereas discrete, episodic, or intermittent point-in-time physiologic measurements (e.g., routine vital signs obtained at discrete clinical encounters) generally would not be considered to constitute a pattern. 

Ultimately, while the updated guidances do not completely overhaul FDA’s pre-existing policies, the changes do provide more room for manufacturers to innovate without incurring regulation by FDA.  However, the regulatory analysis must still be carefully performed, and the product functions and claims monitored by the manufacturer – particularly if updates are made – to ensure that concluding the product is not a device, or falls under enforcement discretion, remains appropriate.